James A. Garfield
1935 - George Gallupís started a weekly column, America Speaks, based on the new technology of public opinion polls. Gallup used a new kind of public opinion polling method called quota sampling.
1936 - Gallup promised newspapers that were paying to use his survey data that if he didnít predict the results of the 1936 election more accurately than the prediction of the Literary Digest, Gallup would refund their money. The Literary Digest forecasted the election with a sample numbering in the millions. Gallupís quota sample included only three thousand people. But Gallupís sample was more representative of the electorate as a whole. The Literary Digest sample, while very large, was also skewedóit included few working class voters.
1948 - Gallupís forecast of the presidential election, like all the other polls, was wrong. Gallup had stopped polling a week before the election. In that last week, a significant number of voters changed their minds and voted to elect President Truman. Also, quota sampling was not perfect and was replaced by random probability sampling.